A Bayesian approach to online performance modeling for database appliances using gaussian models

Muhammad Bilal Sheikh, Umar Farooq Minhas, Omar Zia Khan, Ashraf Aboulnaga, Pascal Poupart, David J. Taylor

Research output: Chapter in Book/Report/Conference proceedingConference contribution

24 Citations (Scopus)

Abstract

In order to meet service level agreements (SLAs) and to maintain peak performance for database management systems (DBMS), database administrators (DBAs) need to implement policies for effective workload scheduling, admission control, and resource provisioning. Accurately predicting response times of DBMS queries is necessary for a DBA to effectively achieve these goals. This task is particularly challenging due to the fact that a database workload typically consists of many concurrently running queries and an accurate model needs to capture their interactions. Additional challenges are introduced when DBMSes are run in dynamic cloud computing environments, where workload, data, and physical resources can change frequently, on-the-fly. Building an efficient and highly accurate online DBMS performance model that is robust in the face of changing workloads, data evolution, and physical resource allocations is still an unsolved problem. In this work, our goal is to build such an online performance model for database appliances using an experiment-driven modeling approach. We use a Bayesian approach and build novel Gaussian models that take into account the interaction among concurrently executing queries and predict response times of individual DBMS queries. A key feature of our modeling approach is that the models can be updated online in response to new queries or data, or changing resource allocations. We experimentally demonstrate that our models are accurate and effective - our best models have an average prediction error of 16.3% in the worst case.

Original languageEnglish
Title of host publicationProceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops
Pages121-130
Number of pages10
DOIs
Publication statusPublished - 15 Jul 2011
Externally publishedYes
Event8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops - Karlsruhe, Germany
Duration: 14 Jun 201118 Jun 2011

Other

Other8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops
CountryGermany
CityKarlsruhe
Period14/6/1118/6/11

Fingerprint

Performance Modeling
Gaussian Model
Bayesian Approach
Workload
Query
Performance Model
Resource Allocation
Response Time
Resource allocation
Service Level Agreement
Resources
Admission Control
Prediction Error
Database Systems
Cloud Computing
Interaction
Modeling
Model
Cloud computing
System Performance

Keywords

  • bayesian networks
  • database appliances
  • gaussian processes
  • machine learning
  • performance modeling

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

Sheikh, M. B., Minhas, U. F., Khan, O. Z., Aboulnaga, A., Poupart, P., & Taylor, D. J. (2011). A Bayesian approach to online performance modeling for database appliances using gaussian models. In Proceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops (pp. 121-130) https://doi.org/10.1145/1998582.1998603

A Bayesian approach to online performance modeling for database appliances using gaussian models. / Sheikh, Muhammad Bilal; Minhas, Umar Farooq; Khan, Omar Zia; Aboulnaga, Ashraf; Poupart, Pascal; Taylor, David J.

Proceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops. 2011. p. 121-130.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sheikh, MB, Minhas, UF, Khan, OZ, Aboulnaga, A, Poupart, P & Taylor, DJ 2011, A Bayesian approach to online performance modeling for database appliances using gaussian models. in Proceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops. pp. 121-130, 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops, Karlsruhe, Germany, 14/6/11. https://doi.org/10.1145/1998582.1998603
Sheikh MB, Minhas UF, Khan OZ, Aboulnaga A, Poupart P, Taylor DJ. A Bayesian approach to online performance modeling for database appliances using gaussian models. In Proceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops. 2011. p. 121-130 https://doi.org/10.1145/1998582.1998603
Sheikh, Muhammad Bilal ; Minhas, Umar Farooq ; Khan, Omar Zia ; Aboulnaga, Ashraf ; Poupart, Pascal ; Taylor, David J. / A Bayesian approach to online performance modeling for database appliances using gaussian models. Proceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops. 2011. pp. 121-130
@inproceedings{d7b1158025584a76a2d4390165c37623,
title = "A Bayesian approach to online performance modeling for database appliances using gaussian models",
abstract = "In order to meet service level agreements (SLAs) and to maintain peak performance for database management systems (DBMS), database administrators (DBAs) need to implement policies for effective workload scheduling, admission control, and resource provisioning. Accurately predicting response times of DBMS queries is necessary for a DBA to effectively achieve these goals. This task is particularly challenging due to the fact that a database workload typically consists of many concurrently running queries and an accurate model needs to capture their interactions. Additional challenges are introduced when DBMSes are run in dynamic cloud computing environments, where workload, data, and physical resources can change frequently, on-the-fly. Building an efficient and highly accurate online DBMS performance model that is robust in the face of changing workloads, data evolution, and physical resource allocations is still an unsolved problem. In this work, our goal is to build such an online performance model for database appliances using an experiment-driven modeling approach. We use a Bayesian approach and build novel Gaussian models that take into account the interaction among concurrently executing queries and predict response times of individual DBMS queries. A key feature of our modeling approach is that the models can be updated online in response to new queries or data, or changing resource allocations. We experimentally demonstrate that our models are accurate and effective - our best models have an average prediction error of 16.3{\%} in the worst case.",
keywords = "bayesian networks, database appliances, gaussian processes, machine learning, performance modeling",
author = "Sheikh, {Muhammad Bilal} and Minhas, {Umar Farooq} and Khan, {Omar Zia} and Ashraf Aboulnaga and Pascal Poupart and Taylor, {David J.}",
year = "2011",
month = "7",
day = "15",
doi = "10.1145/1998582.1998603",
language = "English",
isbn = "9781450306072",
pages = "121--130",
booktitle = "Proceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops",

}

TY - GEN

T1 - A Bayesian approach to online performance modeling for database appliances using gaussian models

AU - Sheikh, Muhammad Bilal

AU - Minhas, Umar Farooq

AU - Khan, Omar Zia

AU - Aboulnaga, Ashraf

AU - Poupart, Pascal

AU - Taylor, David J.

PY - 2011/7/15

Y1 - 2011/7/15

N2 - In order to meet service level agreements (SLAs) and to maintain peak performance for database management systems (DBMS), database administrators (DBAs) need to implement policies for effective workload scheduling, admission control, and resource provisioning. Accurately predicting response times of DBMS queries is necessary for a DBA to effectively achieve these goals. This task is particularly challenging due to the fact that a database workload typically consists of many concurrently running queries and an accurate model needs to capture their interactions. Additional challenges are introduced when DBMSes are run in dynamic cloud computing environments, where workload, data, and physical resources can change frequently, on-the-fly. Building an efficient and highly accurate online DBMS performance model that is robust in the face of changing workloads, data evolution, and physical resource allocations is still an unsolved problem. In this work, our goal is to build such an online performance model for database appliances using an experiment-driven modeling approach. We use a Bayesian approach and build novel Gaussian models that take into account the interaction among concurrently executing queries and predict response times of individual DBMS queries. A key feature of our modeling approach is that the models can be updated online in response to new queries or data, or changing resource allocations. We experimentally demonstrate that our models are accurate and effective - our best models have an average prediction error of 16.3% in the worst case.

AB - In order to meet service level agreements (SLAs) and to maintain peak performance for database management systems (DBMS), database administrators (DBAs) need to implement policies for effective workload scheduling, admission control, and resource provisioning. Accurately predicting response times of DBMS queries is necessary for a DBA to effectively achieve these goals. This task is particularly challenging due to the fact that a database workload typically consists of many concurrently running queries and an accurate model needs to capture their interactions. Additional challenges are introduced when DBMSes are run in dynamic cloud computing environments, where workload, data, and physical resources can change frequently, on-the-fly. Building an efficient and highly accurate online DBMS performance model that is robust in the face of changing workloads, data evolution, and physical resource allocations is still an unsolved problem. In this work, our goal is to build such an online performance model for database appliances using an experiment-driven modeling approach. We use a Bayesian approach and build novel Gaussian models that take into account the interaction among concurrently executing queries and predict response times of individual DBMS queries. A key feature of our modeling approach is that the models can be updated online in response to new queries or data, or changing resource allocations. We experimentally demonstrate that our models are accurate and effective - our best models have an average prediction error of 16.3% in the worst case.

KW - bayesian networks

KW - database appliances

KW - gaussian processes

KW - machine learning

KW - performance modeling

UR - http://www.scopus.com/inward/record.url?scp=79960184485&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79960184485&partnerID=8YFLogxK

U2 - 10.1145/1998582.1998603

DO - 10.1145/1998582.1998603

M3 - Conference contribution

AN - SCOPUS:79960184485

SN - 9781450306072

SP - 121

EP - 130

BT - Proceedings of the 8th ACM International Conference on Autonomic Computing, ICAC 2011 and Co-located Workshops

ER -